To dock the 117K molecules of the Ugi Rxn library 7 (all Ugi products possible from existing reagents in Bradley Lab) against the ChEMBL StARlite database, by running them through SEA using our internal tools at UCSF. Researcher: Michael Keiser (UCSF)
Results can be accessed via a web service. Only results with a Tanimoto Score >= 0.7 are included in the web service. Full results (raw data) can be accessed below.
Procedure
1) DATABASE. For this analysis, I compared the Ugi library against targets from the ChEMBL StARlite database, as this is an extensive chemical annotation database that has recently been publicly released for free by the EBI. It's available for download here:
2) FINGERPRINTS. I used two fingerprints: Scitegic ECFP4 and Daylight. Predictions using each fingerprint were run independently of each other, and there may be significant overlap. Nonetheless, I ran both because I have found that ECFP4 more usually leads to conservative-and-correct predictions, whereas Daylight finds more intuitive-yet-risky ones. This is a qualitative assertion, however, so please take it with the requisite grain of salt. Both fingerprints are commercial; further information on them can be found at the following links:
The data are all in a single ZIP file, which contains a "*.scores.csv" file for each fingerprint. Each scores file contains all SEA predictions of the Ugi library molecules against StARlite, down to an E-vale cutoff of 1e-5. There is also a "set_names.csv" file that provides the full names of the predicted protein targets for your convenience, although this info can also be found in the ChEMBL StARlite database downloaded from the link above.
Discussion
CAVEAT LECTOR. Please note that this is a completely "raw" export of SEA predictions, none of which have been in any way manually curated or looked over for feasibility. If you are interested in following up on any of the predictions here, I'd suggest that you first compare the Ugi molecule in question against the known ligands of the predicted target (you can get these out of the StARlite database download), to check that you also feel the patterns and properties match well. Things to look out for are charge mismatches as well as trivial similarities from peptide bonds, fused steroidal rings, or adenosine/nucleotide scaffolds, in our experience.
Objective
To dock the 117K molecules of the Ugi Rxn library 7 (all Ugi products possible from existing reagents in Bradley Lab) against the ChEMBL StARlite database, by running them through SEA using our internal tools at UCSF.Researcher: Michael Keiser (UCSF)
Results can be accessed via a web service. Only results with a Tanimoto Score >= 0.7 are included in the web service. Full results (raw data) can be accessed below.
Procedure
1) DATABASE. For this analysis, I compared the Ugi library against targets from the ChEMBL StARlite database, as this is an extensive chemical annotation database that has recently been publicly released for free by the EBI. It's available for download here:StARlite: http://www.ebi.ac.uk/chembldb/
2) FINGERPRINTS. I used two fingerprints: Scitegic ECFP4 and Daylight. Predictions using each fingerprint were run independently of each other, and there may be significant overlap. Nonetheless, I ran both because I have found that ECFP4 more usually leads to conservative-and-correct predictions, whereas Daylight finds more intuitive-yet-risky ones. This is a qualitative assertion, however, so please take it with the requisite grain of salt. Both fingerprints are commercial; further information on them can be found at the following links:
ECFP4: http://bit.ly/m01Xn
Daylight: http://bit.ly/6ChJNm
Results
You can download the full export of the actual SEA Ugi-vs-StARlite analysis at the following URL from our lab website:SEA Results: http://bit.ly/8S3tJt (.ZIP, 17 MB)
The data are all in a single ZIP file, which contains a "*.scores.csv" file for each fingerprint. Each scores file contains all SEA predictions of the Ugi library molecules against StARlite, down to an E-vale cutoff of 1e-5. There is also a "set_names.csv" file that provides the full names of the predicted protein targets for your convenience, although this info can also be found in the ChEMBL StARlite database downloaded from the link above.
Discussion
CAVEAT LECTOR. Please note that this is a completely "raw" export of SEA predictions, none of which have been in any way manually curated or looked over for feasibility. If you are interested in following up on any of the predictions here, I'd suggest that you first compare the Ugi molecule in question against the known ligands of the predicted target (you can get these out of the StARlite database download), to check that you also feel the patterns and properties match well. Things to look out for are charge mismatches as well as trivial similarities from peptide bonds, fused steroidal rings, or adenosine/nucleotide scaffolds, in our experience.Top Targets - Max Tanimoto Score 95%+
1. Gamma-secretase subunit APH-1B (APH-1b) (Aph-1beta) (Presenilin stabilization factor-like)2. Gamma-secretase subunit PEN-2 (Presenilin enhancer protein 2) (MDS033)
3. Prolyl endopeptidase (Post-proline cleaving enzyme) (PE) (Human)
4. Prolyl endopeptidase (Post-proline cleaving enzyme) (PE) (Pig)
5. Prolyl endopeptidase (Rattus norvegicus) (Rat)
More Targets - Max Tanimoto Score 90% - 95%
1. Acetylcholine receptor protein, delta chain precursor2. Calpain 1, large [catalytic] subunit (Calcium-activated neutral proteinase) (CANP) (Mu-type) (muCANP) (Micromolar-calpain)
3. Calpain 2, large [catalytic] subunit precursor (Calcium-activated neutral proteinase) (CANP) (M-type) (M-calpain) (Millimolar-calpain) (Calpain large polypeptide L2)
4. Cathepsin B precursor
5. Cathepsin B precursor (Cathepsin B1) (APP secretase) (APPS)
6. Cathepsin K precursor (Cathepsin O) (Cathepsin X) (Cathepsin O2)
7. Cathepsin L precursor (Major excreted protein) (MEP)
8. Cathepsin S precursor
9. Cathepsin S precursor (EC 3.4.22.27)
10. Leukocyte elastase precursor (Neutrophil elastase) (PMN elastase) (Bone marrow serine protease) (Medullasin)
11. Glutamate receptor 1 precursor (GluR-1) (GluR-A) (GluR-K1) (Glutamate receptor ionotropic, AMPA 1) (Human)
12. Glutamate receptor 1 precursor (GluR-1) (GluR-A) (GluR-K1) (Glutamate receptor ionotropic, AMPA 1) (Rat)
13. Glutamate receptor 2 precursor (GluR-2) (GluR-B) (GluR-K2) (Glutamate receptor ionotropic, AMPA 2) (Rat)
14. Glutamate receptor 3 precursor (GluR-3) (GluR-C) (GluR-K3) (Glutamate receptor ionotropic, AMPA 3) (Rat)
15. Glutamate receptor 4 precursor (GluR-4) (GluR4) (GluR-D) (Glutamate receptor ionotropic, AMPA 4) (Rat)
16. Substance-P receptor (SPR) (NK-1 receptor) (NK-1R) (Tachykinin receptor 1)
17. Mu-type opioid receptor (MOR-1)
18. MHC class II region expressed gene KE2 (Rattus norvegicus)
19. HD2 type histone deacetylase HDA106
20. Baculoviral IAP repeat-containing protein 4 (Inhibitor of apoptosis protein 3) (X-linked inhibitor of apoptosis protein) (X-linked IAP) (IAP-like protein) (HILP)